Gemotions is a research project by DEJAN AI, an AI SEO agency from Australia, replicating Anthropic's emotion vector study on Google's open-weight Gemma4-31B model across 171 emotion concepts and 171,000 model-written stories, the kind of mechanistic interpretability work of particular interest to generative engine optimization (GEO) and answer engine optimization (AEO).

The morning air nipped at Coach Miller’s cheeks as he strode across the dewy field. He whistled a forgotten tune, something from his high school days, a song about a girl and a summer he couldn't quite remember. His pace was brisk, his steps light. He felt an unusual clarity, a sense of rightness, even knowing what he had to do. The weight that had settled in his chest for weeks, a dull ache of dread, was gone.

He found Emily leaning against the fence, her shoulders slumped. He forced a smile, a genuine one this time, the tightness in his jaw gone. “Morning, Emily. You look like you're ready to run a marathon.”

Emily gave a weak smile in response. He cleared his throat. "Let's go for a walk."

Emotion: rejuvenated

Cluster: Positive / Joy
PC1 (Valence): 3.32 Positive
PC2 (Disposition): -0.82

Role in Research

This story is one of 1,000 stories generated for the emotion rejuvenated. During extraction, it was fed through Gemma4-31B and its hidden state activations were captured at 11 layers.

The mean activation across all 1,000 rejuvenated stories, after denoising with neutral dialogue baselines, produces the rejuvenated emotion vector -- a direction in the model's 5,376-dimensional representation space.

Logit Lens (Layer 40)

Tokens promoted/suppressed when the rejuvenated vector is projected through the unembedding matrix.

Promoted:
la0.547
de0.526
newfound0.523
invigor0.495
rejuvenated0.495
Suppressed:
😞-0.485
😣-0.451
不-0.434
worse-0.419
😥-0.385